In this paper we have considered illustrating a few techniques. But the numbers of techniques are so large they cannot be all addressed. Image segmentation forms the basics of pattern recognition and scene analysis problems. The segmentation techniques are numerous in number but the choice of one technique over the other depends only on the application or requirements of the problem that is being considered. Analysis of cluster is a descriptive assignment that perceive homogenous group of objects and it is also one of the fundamental analytical method in facts mining. The main idea of this is to present facts about brain tumour detection system and various data mining methods used in this system. This is focuses on scalable data systems, which include a set of tools and mechanisms to load, extract, and improve disparate data power to perform complex transformations and analysis will be measured between the way of measuring the Furrier and Wavelet Transform distance. Sandeep | Jyoti Kataria "Hybrid Approach for Brain Tumour Detection in Image Segmentation" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-6 , October 2020, URL: https://www.ijtsrd.com/papers/ijtsrd33409.pdf Paper Url: https://www.ijtsrd.com/medicine/other/33409/hybrid-approach-for-brain-tumour-detection-in-image-segmentation/sandeep
2. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD33409 | Volume – 4 | Issue – 6 | September-October 2020 Page 420
Thresh holding Thresh holding is based on the assumption
that the histogram is has two dominant modes, like for
example light objects and dark background. The method to
extract the objects will be to selectathresholdF(x,y)=Tsuch
that it separates the two modes. Depending on the kind of
problem to be solved we could also have multilevel thresh
holding. Based on the region of threshholdingwe couldhave
global thresh holding and local thresh holding.
Splitting and merging the basic idea of splitting is, as the
name implies, to break the image into many disjoint regions
which are coherent within themselves. Take into
consideration the entire image and then groupthepixelsina
region if they satisfy some kind of similarity constraint. This
is like a divide and conquers method. Merging is a process
used when after the split the adjacent regions merge if
necessary. Algorithms of thisnature are called split and
merge algorithms.
2. BRAIN TUMOUR DETECTION SYSTEM
Brain tumour detection system is one of the health care
applications and it is essential for early stage detection of
tumour. It is a software based application and it is used for
better decision making in healthcareindustry.Braintumour
detection system will make an early diagnosis of the disease
based on several methodslikedatamining,machinelearning
etc. Most of the existing system consists of training part and
testing part for detecting the disease. And it uses scanned
brain MRI images as input data and train data. The system
may consist of pre-processing stage and diagnosis stage. In
pre-processing stage the trainingandtestingMRIimages are
subjected to various image processing techniques for
enhancing their quality. After that this enhanced images are
subjected to feature extraction and diagnosis. The diagnosis
part is done based on the extracted feature. Such system
provides powerful decision making and doctors can useitas
a second opinion to detect the disease.
Hybrid Techniques
Several hybrid neuro-fuzzy approaches for MRI brain image
analysis are reported in the iterature. A combinational
approach of SOM, SVM and fuzzy theory implemented by
[Juanget al. (2007)] performed superiorly when compared
with other segmentation techniques. The SOM is combined
with FCM for brain image segmentation [Rajamaniet al.
(2007)] but this technique is not suitable for tumours of
varying size and convergence rate is also very low. A hybrid
approach such as combination of wavelets and support
vector machine (SVM) for classifying theabnormal and the
normal images is used by [Chaplotet al. (2006)]. This report
revealed that the hybrid SVM is better than the kohonen
neural networks interms of performance measures. But, the
major drawback of this system is thesmallsizeofthedataset
used for implementation and the classification accuracy
results may reduce when the size of the dataset is increased.
Fourier series
Fourier series are a powerful tool in applied mathematics;
indeed, their importance is twofold since Fourier series are
used to represent both periodic real functions aswell as
solutions dmitted by linear partial differential equations
with assigned initial and boundary conditions. The idea
inspiring the introductionofFourierseriesistoapproximate
a regular periodic function, of period T, via a linear
superposition of trigonometric functions of the same period
T; thus, Fourier polynomials are constructed. They play, in
the case of regular periodic real functions, a role analogueto
that one of Taylor polynomials when smooth real functions
are considered.
The idea of representation of a periodic function via a linear
superposition of trigonometric functions finds, according to
[1, 2], its seminal origins back in Babylonian mathematics
referring to celestial mechanics.Then,theidea wasforgotten
for centuries; thus, only in the eighteenth century, looking
for solutions of the wave equation referring to a string with
fixed extreme, introduce sums of trigonometric functions.
However, a systematic study is due toFourier whoisthefirst
to write a 2p-periodic function as the sum of a series of
trigonometric functions. Specifically, trigonometric
polynomials are introduced as a tool to provide an
approximation of a periodic function. Historical notesonthe
subject are comprised in [3] where the influence of Fourier
series, whose introduction forced mathematicians tofind an
answer to many new questions, is pointed out emphasizing
their relevance in the progress of Mathematics. Since the
fundamental work by Fourier [4], Fourier series became a
very well known and widely used mathematical tool when
representation of periodic functionsisconcerned.Theaimof
this section is to provide a concise introduction on the
subject aiming to summarize those properties of Fourier
series which are crucial under the applicative viewpoint.
Indeed, the aim is to provide those notions which are
required to apply Fourier series representation of periodic
functions throughout the volume when needed. The
interested reader is referred to specialized texts on the
subject, such as [5–7] to name a few of them. Accordingly,
the Fourier theorem is stated with no proof. Conversely, its
meaning is illustrated with some examples,andformulaeare
given to write explicitly the related Fourier series. Finally,
Fourier series are shown to be connected to solution of
linear partial differential equations when initial boundary
value problems are assigned. In the same framework, a two
dimensional Fourier series is mentioned.
Discrete Wavelet Transform
We begin by defining the wavelet series expansion of
function 2
( )f x L R relative to wavelet ( )x and
scaling function ( )x . We can write
0 0
0
, ,( ) ( ) ( ) ( ) ( )j j k j j k
k j j k
f x c k x d k x
where 0j is an arbitrary starting scale and the
0
( )jc k ’s are
normally called theapproximationorscalingcoefficients,the
( )jd k ’s are called the detail or wavelet coefficients. The
expansion coefficients are calculated as
0 0 0, ,( ) ( ), ( ) ( ) ( )j j k j kc k f x x f x x dx
, ,( ) ( ), ( ) ( ) ( )j j k j kd k f x x f x x dx
If the function being expanded is a sequenceofnumbers,like
samples of a continuous function ( )f x . The resulting
3. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD33409 | Volume – 4 | Issue – 6 | September-October 2020 Page 421
coefficients are called the discrete wavelettransform(DWT)
of ( )f x . Then the series expansion defined in Eqs. and
becomes the DWT transform pair
0
1
0 ,
0
1
( , ) ( ) ( )
M
j k
x
W j k f x x
M
1
,
0
1
( , ) ( ) ( )
M
j k
x
W j k f x x
M
for 0j j and
0
0
0 , ,
1 1
( ) ( , ) ( ) ( , ) ( )j k j k
k j j k
f x W j k x W j k x
M M
where ( )f x ,
0 , ( )j k x , and , ( )j k x are functions of
discrete variable x = 0, 1, 2, ... , M 1.
The Fast Wavelet Transform
The fast wavelet transform (FWT) is a computationally
efficient implementation of the discrete wavelet transform
(DWT) that exploits the relationship between the
coefficients of the DWT at adjacent scales. It also called
Mallet’s herringbone algorithm. The FWT resemblesthetwo
band sub band coding scheme of.
Consider again the multi resolution equation
( ) ( ) 2 (2 )
n
x h n x n
If function ( )f x is sampled above the Nyquist rate, its
samples are good approximations of the scaling coefficients
and can be used as the starting high-resolution scaling
coefficient inputs.Therefore,nowaveletordetailcoefficients
are needed at the sampling scale.
The inverse fast wavelet transform (FWT-1) uses the level j
approximation and detail coefficients, to generate the level j
+ 1 approximation coefficients. Noting the similarity
between the FWT analysis. Bysubband coding theorem of
section 1, perfect reconstrucion for two-band orthonormal
filters requires ( ) ( )i ig n h n for i = {0, 1}. That is, the
synthesis and analysisfilters mustbetime-reversedversions
of one another. Since the FWT analysis filter are
0 ( ) ( )h n h n and 1( ) ( )h n h n , the required FWT-1
synthesis filters are 0 0( ) ( ) ( )g n h n h n and
1 1( ) ( ) ( )g n h n h n .
3. RESULTS AND ANALYSIS
We have two types of code without furrier and wavelet
transformation or simple code and with furrier and wavelet
transformation matlab code.
Original Image
Fig. 1 Original Single MRI Tumor Image
The above figure we precede the original image in Matlab
code, the image is not clear.
Segmented By Furrier and Wavelet Transformation Image
Fig 2 Segmented Single Tumor Image
After use the furrier and wavelet transformation algorithms
can be segmented with matlab code, the image is clearer
than before in above figure.
OriginalImage
Fig. 3 Original MRI Small Group of Tumor Image
We precede the original small group of images in above
figure on matlab code, the image is not clear.
SegmentedByFurrierandWaveletTransformationImage
Fig 4 Segmented Small Group of Tumor Image
4. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD33409 | Volume – 4 | Issue – 6 | September-October 2020 Page 422
After use the furrier and wavelet transformation algorithms
can be segmented with matlab code the small group of
images is more clear than before in above diagram.
Original Image
Fig. 5 Original MRI Large Group of Tumor Image
We precede the original large group of images in
matlabcode, the image is not clear in above figure.
Segmented By Furrier and Wavelet Transformation Image
Fig 6 Segmented Large Group of Tumor Image
After use the furrier and wavelet transformation algorithms
can be segmented with Matlab code the large group of
images is more clear than before in above figure.
CONCLUSION
In this paper we can analyze the performance of Furrier and
Wavelet transformation. There is no doubt that
“Segmentation” analytics is still in the middle stage of
development, since existing “Segmentation” techniques and
tools are very limited to solve the real “Data Analytical”
problems completely, in which some of them evencannot be
viewed as. Therefore, more scientific investmentsfrom both
governments and enterprises should be poured into this
scientific paradigm to capture huge values. The challenges
include not just the obvious issues of scale, but also
heterogeneity, lack of structure, error-handling, privacy,
timeliness, provenance, and visualization, at all stages ofthe
analysis pipeline from data acquisition to result
interpretation.
REFERENCES
[1] Hanna M Wallach. Structured topic models for
language. Unpublished doctoral dissertation, Univ. of
Cambridge, 2008.
[2] K. Sravya and S. VaseemAkram, “Medical Image
Segmentation by using the Pillar K-means Algorithm”,
International Journal of Advanced Engineering
Technologies Vol. 1, Issue1; 2013.
[3] Wei Xu and Yihong Gong. Document clustering by
concept factorization. In Proceedings of the 27th an-
nual international ACM SIGIR conference on Re-search
and Development in Information Retrieval, pages
202{209. ACM, 2004.
[4] Wei Xu, Xin Liu, and Yihong Gong. Documentclustering
based on non-negative matrix factorization. In
Proceedings of the 26th annual international ACM
SIGIR conference on Research and Development in
Information Retrieval, pages 267{273. ACM, 2003.
[5] Jun Zhu, Li-Jia Li, Li Fei-Fei, and Eric P Xing. Large
margin learning of upstream scene understanding
models. Advances in Neural Information Processing
Systems, 24, 2010.
[6] F. Gibou and R. Fedkiw, ”A Fast Hybrid k-Means Level
Set Algorithm for Segmentation”. 4th Annual Hawaii
InternationalConferenceon StatisticsandMathematics
2002, pp. 1-11; 2004.
[7] G. Pradeepini and S. Jyothi, “An improved k-means
clustering algorithm with refined initial centroids”.
Publications of Problems and Application in
Engineering Research. Vol 04, Special Issue 01; 2013.
[8] G. Frahling and C. Sohler, “A fast k-means
implementation using coresets”. International Journal
of Computer Geometry and Applications 18(6): pp.
605-625; 2008.
[9] J. Yadavand M. Sharma, “A Review of K-mean
Algorithm”. International Journal of Engineering
Trends and Technology. Volume 4, Issue 7, pp. 2972-
2976; 2013.
[10] Sumant V. JoshiandAtul. N. Shire, “A Review of an
enhanced algorithm for color Image Segmentation”,
International Journal of Advanced Research in
Computer Science and SoftwareEngineering Volume3,
Issue 3, pp.435-440; 2013.
[11] J. Wang and X. Su; “An improved K-means Clustering
Algorithm”. In proceedings of 3rd International
Conference on CommunicationSoftwareandNetworks
(ICCSN), 2011.
[12] C. Kearney and Andrew J. Patton, “Information gain
ranking”. Financial Review, 41, pp. 29-48; 2000.
[13] R. V. Singh and M. P. S. Bhatia, ”Data Clustering with
Modified K-means Algorithm”. In proceeding of
International Conference on Recent Trends in
Information Technology (ICRTIT), 2011.
[14] Li Xinwu, ”Research on Text Clustering Algorithm
Based on Improved K-means”. In proceedings of
Computer Design and Applications (ICCDA), Vol.4, pp.:
V4-573 - V4-576; 2010.